3
$\begingroup$

Have I understood it correctly if the standard way to calculate implied correlation is the Gaussian Copula model where we:

  • Calibrate the underlying portfolio to get a homogenous default probability for all assets.
  • Calibrate the correlation parameter alpha to each tranche in the CDO.

I am a bit confused since the most articles i've read do not comment on the default probability, but only the correlation parameter and it seems counterintuitive to ignore one parameter.

p.s. it seems that the idea of implied correlation has been abandoned after the crash in 2008, what is a viable alternative?

| improve this question | |
$\endgroup$
  • $\begingroup$ see here. $\endgroup$ – will Aug 4 '17 at 16:35
  • $\begingroup$ For the gaussian copula with a homogenous portfolio, I get the conditional default probability: $$ \mathbb{Q}(\tau\leq T|W_t=w)=\phi\left(\frac{C-\rho w}{\sqrt{1-\rho^2}}\right) $$ Here we need two parameters C and $\rho$. However in the pricing model presented by Brian B., he only demands one parameter $\rho$. What assumptions are made to make C disappear? @will $\endgroup$ – Upsimus Aug 6 '17 at 8:38
  • $\begingroup$ I understand that we assume the hazard rate to be a constant, but what is a reasonable value for such a constant? $\endgroup$ – Upsimus Aug 7 '17 at 11:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.